Drug Discovery in 90 Seconds: Build and Run a Screen in Talindrew's AI Workspace

About this video

A whole drug-discovery screen, built and run in about a minute in Talindrew's AI workspace. Start from an empty canvas and paste a target protein (myoglobin, 154 residues), then add each stage as a node: fold, pockets, compounds, docking, ADMET, pose checks, off-target selectivity, ranking and design. Each node connects itself to the stages it reads from. Run it, and every node shows what it found: Imatinib docks best, but once all the evidence is weighed in, Celecoxib ranks first, and the design stage proposes 184 analogs for round two. The film replays a finished screen; a live run on GPUs takes about 10 minutes.

What you will learn

• Build a virtual screening pipeline from a target protein, one stage at a time
• Fold the target, find binding pockets and dock a compound library
• Weigh docking against ADMET, pose checks and off-target selectivity in one ranking
• Propose analogs of the best hits for the next round